The protection from harmful aerosols afforded by negative pressure air- purifying respirators depends on both the behavior of filters and valves and facepiece fit. It has been hypothesized that the latter is the most significant factor in a respirators protection. However, respirator valve failure may add significantly to a respirators reduced protection. No work has been done to demonstrate valve behavior as stressed with cyclic flow over time. This research will examine aerosol penetration and leak flow through exhalation and inhalation valves from negative pressure air-purifying respirators. New valves from several manufacturers will be stressed with an eight hour cyclic flow at two workrates. Penetration of five particle sizes between 0.1 and 1 mu-m diameter will be measured periodically during each test. At the same time, valve leak flow under static pressure will be determined, in a manner similar to that presently used by the National Institute for Occupational Safety and Health for testing respirator exhalation valves. Used valves will be obtained from respirators which have been worn in industrial work situations and similar penetration and leak flow measurements will be made. Workrate and usage time will be estimated by observing and interviewing the respirator wearers, and history of use and maintenance will be obtained from employer's records. A model to describe particle size-related penetration for a given workrate and leak flow will be developed using measurements of leak area and valve dimensions. Predictions using this model will be compared with results obtained at a different workrate. Valves are important elements of a respirator and their partial failure could severely compromise the protection offered. These experiments will identify initial aerosol penetration characteristics of respirator valves, as well as changes which may occur due to stress over time. Effects of valve model (manufacturer), workrate, and maintenance program will be explored to gain an understanding of the range of possible valve efficacy. Results of these experiments will help in the development of better valve test methods, the design of better valves, and the identification' of appropriate valve usage time and maintenance. In addition, these results will assist in identifying the influence of valves on the protection offered by a respirator over time and usage.